[american institute of aeronautics and astronautics 46th aiaa aerospace sciences meeting and exhibit...
TRANSCRIPT
1
American Institute of Aeronautics and Astronautics
1Compressor Aerodynamic Design Engineer, 6225A W Sam Houston Pkwy N, Houston, TX 77041
2Professor, Dept. of Aerospace Engineering, 223 Hammond Building, University Park, PA 16802
Pressure Resolution of a PSP Based Measurement
System with Non-Linear Intensity Response
Tarun Dhall1
Wood Group Light Industrial Turbines (USA) Inc., Houston, TX 77041
and
Cengiz Camci2
Turbomachinery Aero-Heat Transfer Laboratory, Department of Aerospace Engineering, The
Pennsylvania State University, University Park, PA 16802
An experimental study has been conducted at The Pennsylvania State
University’s Turbomachinery Aero-Heat Transfer Laboratory to develop an
automated facility for calibration of Pressure Sensitive Paints (PSP). This paper
deals with the calibration results of a particular PSP sample and their associated
implications. Experimental setup of the facility and the calibration procedure have
been briefly discussed. Results of this study, presented here in detail, reveal a
unique character of the intensity ratio – pressure ratio relationship for the sample
PSP. Rather than following the conventional linear pattern, this relationship was
found to have a highly nonlinear character which was approximated with an
exponential function using curve-fitting analysis. Repeated tests were conducted to
confirm this observation. Such uniqueness of this relationship has a significant
impact on not only the sensitivity range but also the pressure resolution capability of
the paint. Since the intensity resolution of a PSP based measurement system is fixed
(depending on the imaging device used), an exponential relation between intensity
and pressure will lead to a variation in the corresponding pressure resolution over
the whole range of pressures. In the present case, where a 10-bit CCD camera was
used, such a variation was observed to be following a fourth-order polynomial
function. Further, a simulated analysis for a 16-bit camera showed a significant
improvement in pressure resolution for the same paint, although it still exhibited
distinct variability over the complete pressure range. Finally, an error analysis was
carried out over the different tests conducted as part of this study. Pressure values
were estimated using the calibration equation and recorded intensity, and compared
against measured values. It was observed that, in general, rms error increased as the
pressure resolution of the paint decreased.
I. Introduction
RESSURE sensitive paint (PSP) has been fast gaining recognition as an important experimental tool
for optically and quantitatively measuring pressure. Due to the associated ease of setting up the
complete measurement system and the amount of useful data that it is capable of providing, the PSP
technology is seeing applications in both supersonic as well as subsonic flow regimes. However, in all its
applications, the fundamental basis of this technology remains the same. Reflected intensity signals from
the paint applied to the model surface are recorded at wind-on and wind-off conditions using a CCD camera
and are subsequently converted into pressure values using a pre-determined calibration relation for that
specific paint. Such calibration relation is commonly known as the Stern-Volmer equation and relates
intensity ratio and pressure ratio of the known (wind-off) and the unknown (wind-on) conditions. In most
cases, the Stern-Volmer equation is found to be almost linear with some temperature dependence which can
P
46th AIAA Aerospace Sciences Meeting and Exhibit7 - 10 January 2008, Reno, Nevada
AIAA 2008-279
Copyright © 2008 by the American Institute of Aeronautics and Astronautics, Inc. All rights reserved.
2
American Institute of Aeronautics and Astronautics
be suitably overcome by keeping the temperature constant for the reference and the measurement condition.
However, for certain paint formulations the calibration relation does not follow a linear trend. This property
of the PSP, in such cases, gains significance since it directly affects the applicable pressure range for the
paint and its corresponding pressure resolution.
A better understanding of the PSP behavior and the Stern-Volmer equation can be had by looking at the
underlying physical processes which cause the PSP to respond towards any change in pressure in its
surroundings. Navarra1 provides a comprehensive description of such physical processes followed by their
mathematical analysis. In a similar approach, Ingle and Crouch2 and Dhall
3 present their analyses based on
the luminescence quantum yield (or efficiency), φL defined as the ratio of the luminescence radiant power
to the absorbed radiant power where the radiant powers are expressed in photons per second. From its
definition, φL will be dependant on the rate of both emission (luminescence) as well as absorption of
photons. An elaborate treatment of these first-order processes has also been provided by Mosharov et al4
and Bell et al5 where luminescence has been analyzed considering the combined effects of fluorescence and
phosphorescence. Using a few simplifying assumptions based on the fact that the rate of a first order
reaction depends upon the concentration of the constituents, all the mathematical analyses reach at the basic
form of the Stern-Volmer equation (Eq. 1).
where 0
Fφ and Fφ are the fluorescence quantum efficiencies in absence and presence of the quencher,
respectively, Kq is the Stern-Volmer constant representing a combination of first-order rate constants of the
different internal processes taking place including fluorescence, and [Q] is the quencher concentration
which is responsible for the quenching of PSP fluorescence. Since the luminescence quantum efficiency is
defined in terms of photons per second, it can be directly related to the intensity of radiations (I), yielding:
Using Henry’s Law to relate quencher concentration to pressure and acknowledging oxygen as the
primary quencher in case of PSP fluorescence, the concentration of oxygen, [O2] can be written as:
where P is the air pressure and S is a temperature dependant Henry’s Law coefficient. Further, in order to
eliminate 0I which is almost impossible to estimate, a ratio of the Stern-Volmer equation for two different
conditions is taken where the reference condition represents the known condition. Hence, the final
simplified form of the Stern-Volmer equation can be expressed as:
A0 and A1 are known as the Stern-Volmer coefficients and since these are based on the definition of S,
they too exhibit temperature dependency. Torgerson6 provides definitions for both these coefficients as
follows:
[ ]QK q
F
F +=10
φφ
(1)
[ ]QKI
Iq+=1
0
(2)
[ ] PSO .2 = (3)
ref
ref
P
PAA
I
I10 += (4)
( )
−+=
ref
ref
ref
nr
T
TT
RT
EaTA 100 (5)
3
American Institute of Aeronautics and Astronautics
where Enr and Ep represent the non radiative and polymer (binder) activation energies, R is the interaction
distance and, a0, a1 are constants. An essential conclusion that can be drawn from Eq. 5 and Eq. 6 is that the
temperature dependency of these constants can be effectively cancelled out by keeping refTT = .
Equation 4 is the most common form of Stern-Volmer equation that is used for PSP measurements due
to linearity of the relationship between relative intensity II ref / and relative pressure refPP / . However,
in case of some special coatings and under certain pressure and temperature conditions, this relation may no
longer remain linear. As explained by McLachlan and Bell7, a more general form of Henry’s law is
required to explain PSP behavior in these cases. Accordingly the Henry’s law coefficient, S is modified to
include the effects of not only temperature abut also pressure. This leads to a co-dependence of constants in
Eq. 4 on pressure as well as temperature. In order to simplify the complex character of these constants, a
polynomial expansion in pressure is used and the resultant, more general form the Stern-Volmer equation is
written as Eq. 7. It is important to note that when used in this form, the Stern-Volmer equation retains its
temperature dependency but attains a certain non-linear characteristic.
In case of a PSP measurement system, the pressure resolution of the system essentially depends on the
intensity resolving capability of the camera used. Using the paint’s calibration relation, intensity
information recorded for each camera pixel is converted to pressure values. Depending upon the bit depth
of the camera (say N), reflected luminescence signal from the paint is recorded in a number of divisions
(2N). Each division corresponds to a particular intensity value and hence a particular pressure value.
Difference between any two consecutive pressure levels, then, becomes the least decipherable pressure
difference and hence the pressure resolution of the PSP system. However, in case a non-linear calibration
relation, a constant intensity resolution of the camera will lead to a non-linear pressure resolution. This
characteristic should be an important consideration while making measurements using the PSP system
since a varying pressure resolution restricts the sensitivity and applicability range of the system.
A similar observation was made while calibrating a sample PSP using a calibration facility developed at
the Turbomachinery Aero-Heat Transfer Laboratory at the Pennsylvania State University. This paper
discusses the results of the calibration testing and focuses on their implications. The complete calibration
process has been briefly explained followed by the results and their analysis.
II. Calibration Facility
Primary objective of this research study was to develop a PSP calibration system which could be used
for accurately determining the correlation between a paint sample’s luminescent response and the applied
pressure. Calibration facility developed as part of this study has been schematically illustrated in Figure 1.
PSP sample to be tested was applied to a base plate kept inside an enclosed calibration chamber. In order to
provide optical access to the paint, the face plate of the chamber was made out of thick plexiglass. A
pressure pump and a thermoelectric cooler were used to control pressure and temperature inside the
chamber, respectively. Excitation illumination for the PSP was provided from a high intensity Xenon-arc
lamp and the reflected luminescence signals were captured using a 10-bit, 1.3 million (1280 x 1024) pixels
scientific grade CCD camera. In case of PSP testing, it is important to ensure that out of the complete
spectrum of the excitation light, only the appropriate frequency radiations reach the paint. Similarly, the
camera should be setup such that it captures radiations generated by the paint and no reflected part of the
incident light. For this purpose, a blue colored bandpass filter (440±5 nm) and a red colored longpass filter
(600±5 nm) were used with the lamp and the camera respectively.
( )
−+=
ref
ref
ref
p
T
TT
RT
EaTA 111 (6)
( ) ( ) ( ) L+
++=
2
210
refref
ref
P
PTA
P
PTATA
I
I
(7)
4
American Institute of Aeronautics and Astronautics
The paint sample used for this work was provided by the Innovative Scientific Solutions, Inc. (ISSI),
Dayton, Ohio. The calibration procedure was carried out by varying the pressure inside the chamber in
small steps from 5.5 psia to 21 psia (approx.) and recording the florescence signal from the paint at each
step. Since, PSP response is known to be temperature dependant, each pressure cycle was repeated at
different temperatures ranging between 18-40°C. In order to increase the signal-to-noise ratio (SNR), for
each test six continuous frames from the camera were recorded and averaged. Following this, difference
between intensity values for pixels corresponding to a certain region of interest in the averaged image and
the background (dark) image are calculated and finally ratioed with difference of the image at atmospheric
pressure (reference pressure) and the background image to obtain the intensity ratio for the tested pressure
ratio (P/Patm). This intensity ratio – pressure ratio data is evaluated analytically to yield a mathematical
relation which serves as the calibration equation for the paint.
III. Calibration Results
Based on the procedure outlined above, a number of tests were carried out for the particular PSP
sample. Although the major goal of these tests was to develop a mathematical formulation to relate the
intensity response of PSP sample to test pressures, certain other essential aspects pertinent to PSP
measurements like photodegradation, temperature effects etc. were also analyzed.
The camera recorded the response data from the paint sample by dividing the complete intensity range
(from black to white light) in 210 (=1024) divisions. Hence, the intensity observations from the image
acquisition system were in terms of numbers ranging from 0 to 1024. Although, due to reflection of light
and reduction of light intensity caused by the use of multiple filters, intensity readings were generally low.
Each test was initiated by recording the paint response at atmospheric pressure. This was followed by
changing the pressure and recording the corresponding response. In order to have a precise and detailed
calibration plot, pressure was changed in small steps (~0.1 psi). Data from the image acquisition system
was continuously passed on to the image processing software especially designed for this study for storage
and further processing. This software focused on a certain region of interest (ROI) in every image and
calculated the intensity ratio between the particular pressure image and the atmospheric pressure image on
a pixel-by-pixel basis, in effect creating a ‘ratioed’ image. Since all intensity ratios in this ratioed image
corresponded to the same pressure, their values were essentially the same. However, in order to be
statistically correct, a histogram of all ratio values was plotted for every ratioed image and the value
occurring the maximum number of times was selected as the final intensity ratio corresponding to that
pressure. As a graphical representation of the results, a plot between all the pressure ratios (P/Patm) and
corresponding intensity ratios (Iatm/I) was plotted. This data was then analyzed through a curve-fitting
routine to generate a mathematical relationship and hence provide the calibration equation. Further, in order
to insure the repeatability and stability of results, several similar tests were conducted.
Figure 1. Schematic of the calibration facility
5
American Institute of Aeronautics and Astronautics
Figure 2 shows the results of a particular calibration test in which each point represents the data
obtained at separate pressures. Also shown in the image, is an exponential plot which was obtained by
fitting the data to an analytical curve. This plot has been mathematically expressed in Eq. 8. It can clearly
be seen that the data fits to the plot extremely well, especially for vacuum pressures. However, a small
degree of scatter is also evident for higher pressures which can be attributed to the lack of a sophisticated
control over generation of higher pressures in the calibration chamber and poor pressure resolution of the
PSP for this range. Calibration results from other tests were also analyzed and corresponding plots have
been shown in Figure 3. A similar exponential dependence exhibited by all the tests confirmed the
repeatability and stability of the calibration equation.
where,
a = 0.9814
b = 0.03894
c = 1.799
d = 4.52
( ) ( )xdcxbay .exp..exp. −−= (8)
Figure 3. Repeatability of Results
Figure 2. Calibration Results
6
American Institute of Aeronautics and Astronautics
Such an observance does not conform to the conventional and more commonly used linear form of
Stern-Volmer equation relating intensity ratio and pressure ratio. This fact is further supported by Figure 4,
which extrapolates the results of previous calibration test to both lower as well as higher pressures. The
change in PSP response at higher pressures may be attributed to the phenomenon of luminescence
saturation of emitted intensity which implies that if at a given pressure almost all the lumiphores present in
the PSP layer are already being quenched, any further increase in pressure would not lead to a consequent
increase in emitted intensity resulting in a constant or uniform response.
PSP is known to exhibit temperature dependence and in order to include the effect of temperature
variations in the calibration equation, tests were also conducted at different temperatures. Figure 5 shows
the results of three such tests conducted on the PSP sample for the same range of pressures. However,
while analyzing and presenting the results care was taken to identify the reference point at the atmospheric
pressure of the temperature at which the test was conducted. This step, in effect, eliminated the dependence
of the calibration equation on temperature.
Figure 4. Extrapolation of Results
Figure 5. Temperature Effects
7
American Institute of Aeronautics and Astronautics
IV. Implications of Non-Linear Calibration Equation
A non-linear calibration equation will have a significant impact on the sensitivity and the pressure
resolution capability of a measurement system designed using such a PSP. It is clear from the calibration
plots that this PSP exhibits an enhanced sensitivity towards measuring vacuum pressures as compared to
higher pressures. This fact puts limits to its range of measurable pressures and hence the overall
applicability of the system.
The most important aspect of any PSP measurement system is its ability to yield spatially continuous
and accurate pressure maps over a test surface. However, since the measurements are based on capturing
intensity response of the paint and all the imaging devices have a definite limit to their intensity resolution,
pressure distribution can also be mapped only with a certain minimum resolution. This minimum pressure
is called the pressure resolution of the PSP measurement system. If, as in the current study, a 10-bit CCD
camera is used for the measurements, the total number of intensity divisions available would be 210
(=1024). Hence, the pressure resolution of that system would be the change in pressure corresponding to a
unit change in intensity. This leads to a significant implication in reference to the non-linear calibration
equation which has been shown in Figure 6 in terms of intensity vs. pressure instead of corresponding
ratios. A unit change in intensity (∆I) can lead to different changes in pressure (∆P1 and ∆P2) depending
upon the location of the point on the plot resulting in a variation of the pressure resolution of the system.
In order to further investigate this variation of resolution, an analysis was carried out where using the
previously obtained calibration equation change in pressure was calculated for a unit change in intensity for
every intensity test point in the range of the previously conducted tests. The reference point data used for
calculating intensity values from corresponding ratios was taken from the test. Owing to its transcendental
nature, the calibration equation was solved numerically at every point to calculate pressure values from
intensity data. The results of this analysis have been presented in Figure 8 where resolution capability of
the system has been plotted against corresponding pressure values. This data was found to be following a
polynomial curve of the fourth order (Eq. 9).
where
a1 = -2.423e-005
a2 = 0.001171
a3 = -0.01514
a4 = 0.08969
a5 = -0.1916
( ) ( ) ( ) ( ) ( ) 54
2
3
3
2
4
1 .... aPaPaPaPaPR ++++= (9)
Figure 6. Variation of Pressure Resolution
8
American Institute of Aeronautics and Astronautics
Two aspects are clearly evident from the plot. Firstly, the pressure resolution is not constant and varies
along the range of pressure. Secondly, the resolution is only good for vacuum pressures and deteriorates
steadily with increase in pressure to almost unacceptable values for higher pressures. Further analysis of
this variation of resolution was carried out by simulating a 16-bit intensity capturing device. In this case,
the intensity readings from before were simply multiplied by a factor of 64 (=2(16-10)
). Pressure resolution
was again obtained using the same numerical procedure and the results, shown in Figure 7, exhibit a similar
pattern of variation of resolution. However, a significant difference is evident in the scale of its values
which is approximately 64 times enhanced. This leads to the conclusion that a CCD camera with higher bit
depth is capable of providing improved pressure resolution for this PSP but it will still suffer from a
definite pattern of variation.
Finally, an error analysis was carried out over all the tests conducted in this study. Using the calibration
equation and recorded intensity values, corresponding pressure values were estimated. Difference between
estimated and recorded pressure values was noted down as the error. Figure 9 shows how the error level
varied over the whole range of tested pressures. As can be seen from the plot, vacuum pressures can be
predicted by the PSP sample more accurately and with less error than higher pressures. Rms error obtained
from the distribution was 6.867 KPa, however, the same value for just the vacuum pressures was 1.674
KPa. This supports previous findings of this study which point towards inability of the paint to predict
Figure 8. Variation of Pressure Resolution 10 bit
Figure 7. Variation of Pressure Resolution 16 bit
9
American Institute of Aeronautics and Astronautics
higher pressures with a greater degree of certainty because of low pressure resolution. Further, owing to
greater sensitivity and enhanced pressure resolution, relatively low error levels were observed for the
vacuum pressures.
V. Conclusions
Application and use of pressure sensitive paint (PSP) depends essentially on the calibration relation of
that PSP. Conventionally, it has always been assumed that this calibration relation is basically a linear
relation between the intensity and pressure ratios. However, by means of an automated calibration facility,
this study revealed a highly non-linear calibration equation for a particular PSP sample. Greatest
implication of such non-linearity was observed on the sensitivity and pressure resolution of the paint which
limits the applicability range of this particular sample.
In conclusion, this paper emphasizes the importance of carefully calibrating the PSP before any
application and setting the goals of that application in accordance to the sensitivity and pressure resolution
of the paint.
References 1Navarra, K.R., “Development of the Pressure-Sensitive Paint Technique for Advanced Turbomachinery
Applications”, M.S. Thesis, April 1997, Mechanical Engineering Department, Virginia Polytechnic Institute and State
University, Blacksburg, VA. 2Ingle, J. D. Jr., and Crouch, S. R., 1988, “Spectrochemical Analysis”, Prentice Hall, New Jersey, Ch. 12, p 338-
351. 3Dhall, T., “Development and Implementation of Pressure Sensitive Paint for Aerodynamic Measurements”, M.S.
Thesis, August 2007, Department of Aerospace Engineering, The Pennsylvania State University, University Park, PA. 4Mosharov, V., Radchenko, V., and Fonov, S., “Luminescent Pressure Sensors in Aerodynamic Experiment”, 1997,
Cent. Aerohydrodyn. Inst. (TsAGI). CWA Int. Corp. 151 pp. 5Bell, J.H., Schairer, E.T., Hand, L.A., and Mehta, R.D., “Surface Pressure Measurements Using Luminescent
Coatings”, 2001, Annual Review of Fluid Mechanics, 33, p 155–206. 6Torgerson, S.D., “Pressure Sensitive Paints for Aerodynamic Applications”, 1997, M.S. Thesis, School of
Aeronautics and Astronautics, Purdue University, West Lafayette, IN. 7McLachlan, B. G., Bell, J. H., “Pressure-Sensitive Paint in Aerodynamic Testing”, 1995, Experimental Thermal
and Fluid Science, 10, p 470-485.
Figure 9. Error Analysis